Protonation and Substitution Studies on Cyclopenta [c

Protonation and Substitution Studies on Cyclopenta [c...
0 downloads 0 Views 1MB Size
708

ARTHURG . ANDERSON,J R . ,

AND

The ether layer was washed with a saturated sodium bicarbonate solution and water, dried over sodium sulfate, and evaporated t o dryness. Crystallization from ether yielded XXVIII (88 mg., m.p. 180-182", 68y0). A sample was recrystallized for analysis

[CONTRIBUTION FROM

THE

WILLIAMF. HARRISON

Vol. 86

from ether; n1.p. 183-185', [ ~ ] Z ' D +119" ( c 0.21, CHCII), 240 mw (15,400); hL!F'3 5.85, 6.00, 6.18 p . Anal. Calcd. for C18H260~:C, 79.68; H, 9.15. Found: C , 79.51; H, 9.42.

DEPARTMEST OF CHEMISTRY, UNIVERSITY OF N'ASHINGTOS, SEATTLE, IVASH ]

Protonation and Substitution Studies on Cyclopenta [clthiapyran and 2-Phenyl-2-pyrindinel BY

-d

AND WILLIAM F.HARRISON4 RECEIVED h U G U S T 31, 1963

ARTHURG . ANDERSON,J R . ,

The behavior of the compounds in the title, which are iso-rr-electronic hetero-analogs of azulene, with certain acids and bases is reported The ultraviolet and n.m.r. spectra and comparisons with azulene indicate the probable positions of protonation. The' successful electrophilic substitution of cyclopenta[c]thiapyran by a number of reagents, the positions of substitution as indicated by the n.tn r. spectra of the products, the effects of different substituents on the visible spectrum, and some reactions involving nucleophilic reagents are described and discussed

The syntheses of the first simple *-excessive heterocyclic analogs of azulene, cyclopenta [clthiapyran ( l ) , and 2-phenyl-2-pyrindine ( 2 ) , and a comparison of some of their physical properties with those of azulene were described in the previous paper.j The similarities observed in the comparison led to the investigation of some of the chemical properties of the heterocycles and t h e present paper deals with results of these studies. Behavior with Acids and Bases.--At the time we began these experiments, the basicity of azulene was well known and theoretical calculations, considerations of the spectral and proton magnetic resonance changes t h a t occurred upon reaction with strong acids, and the analogy of electrophilic substitution provided strong evidence t h a t protonation occurred in the l-position.6 In the early n.m.r. work by Frey,' however, the absorptions attributed to the 2- and 3-protons were not separately resolved and the assignment of the highest field absorption to water impurity appeared to be incorrect. Also the infrared spectral data of Bauder and Gunthard8 on dideuterioazulene did not seem to be really definitive with regard to this question. Therefore the proton magnetic resonance spectra of azulene in both 98% sulfuric acid and 7OCr, deuterated 98% sulfuric acid were examined. The results (Fig. 1) obtained were clearly consistent only with complete, reversible protonation a t the 1- (3.) position. Except t h a t no coupling between the 1,l- and the 2- or 3hydrogens was detectable, the spectrum in sulfuric acid was very similar to t h a t in trifluoroacetic acid subsequently reported by Danyluk and S ~ h n e i d e r . ~ The proton exchange rate in the former solvent is thus appreciably the faster and the mean life-time of the ion is only a fraction of a second. The spectrum in deuterated acid did not change after 15 min. and showed deuteration a t both the 1- and 3-positions. These results plus the earlier analysis of the n.m.r. spectrum of cyclopenta [clthiapyranj provided a basis for the interpretation of the spectrum of this compound in acid (Fig. 2). The lack of symmetry in the molecule served to separate the absorptions for most of the dif11) Supported in part by a grant from the National Science Foundation. ( 2 ) A portion of this work was reported in a preliminary communication A. G Anderson, J r , a n d W F Harrison, Trivahrdron I.ellevr. No.2,11 (1960). ( 3 ) From the P h D thesis of William F Harrison ( 4 ) Standard Oil of California Fellow, summer, I B R 8 , National Science Foundation Predoctoral Fellow, 19.59-1960 (5) A G . Anderson, Jr , W F Harrison. and K G Anderson. J . A n i C h - m . Soc , 86, 3 4 4 8 (IUS:60yGsulfuric acid, and also to 5yo ethanolic potassium hydroxide. At intermediate acidities as in dilute sulfuric acid and in glacial acetic acid, however, decomposition rapidly occurred. The apparent explanation lies in the formation of some of the conjugate acid of the heterocyclic base in the presence of a sufficiently reactive nucleophile. I n the decomposition which ensues, the latter could attack the conjugate acid at a position adjacent t o the hetero-atom, with subsequent ring scission a probable result. The hetero-pseudoazulenes were also unstable to chromatography on Florisil, neutral alumina, and silica ge1.16 The colorless to pale yellow appearance and ultraviolet spectra of solutions of azulene and 1 in concentrated sulfuric acid (Fig. 6) indicated that these were essentially completely in the protonated form. In acetic acid the initial ultraviolet and visible spectra showed the compounds t o be present primarily as the free bases. In contrast, the initial spectrum of 2 was essentially the same in sulfuric (Fig. 6) and acetic acids. Thus azulene and 1 are appreciably weaker bases than acetate ion, and 2 is an appreciably stronger base than acetate ion. Electrophilic Substitution.-The observed basicity of 1 and 2 with proton exchange a t the 5- and 7-positions enhanced the likelihood t h a t the similarity of these compounds to azulene would extend t o electrophilic substitution a t these positions. The relative quanti(13) G. V Boyd, Chem I n d . (London), 37, 12-14 ( l 9 5 7 ) , J . C h r m Soc , 1978 (1958) 114) G V Boyd, ibid., .55 (1959). (1.5) A quite low yield of cyclopenta[c]thiapyran was recovered when a solution o f t h e conjugate acid of this in sulfuric acid was added dropwise to a slurry of aqueous potassium hydroxide and ether (16) An attempt t o capture a possible fulvene aldehyde intermediate decomposition product in the ethereal eluate from chromatography on d i C a gel afforded a brown oil which gave a precipitate with 2,4-dinitrophenylhydrazine b u t could not be further identified.

711

CYCLOPENTA [CITHIAPYRANA N D 2-PHENYL-2-PYRINDINE

Feb. 20, 1964

i

,

...............L -_..--. -. A

.

.

*

...... 4

A.

___

-,.-

.

.

2.96

-.

,-.

. . .

L . . ._ ..............

Fig 7 --N tn r spectrum of 5,7-dichlorocyclopenta[c]thiapyran in partially deuterated chloroform with tetramethylsilane as the internal reference

ties of the substances a t hand and the absence of a second aromatic ring in 1 caused the selection of this compound for these studies. To avoid possible difficulties in the separation of isomeric monosubstitution products, attention was initially directed for the most part t o reagents and conditions which had been found t o give disubstituted products with azulene. Treatment of 1 with two equivalents of N-chloroor X-bromosuccinimide a t room temperature formed a dichloro (667,) or a dibromo (50%) derivative, respectively. An attempt t o effect trichlorination gave only a green polymer-like material. The ultraviolet and visible absorption spectra of the dihalo compounds closely resembled those of 1. Reaction of 1 with excess thiocyanogen gave a 487, yield of a dithiocyano derivative. U-hen the product from the treatment of 1 with mercuric acetate (probably the diacetoxymercuri derivative) was allowed to react with thiocyanogen, a compound having the same ultraviolet and visible spectrum as the dithiocyano derivative was formed in very low yield.17 Acetylation of I with excess acetic anhydride in the presence of stannic chloride gave a low (9%) yield of a diacetyl derivative.I8 Evidence for the structures of these products as 5,7disubstituted cyclopenta [clthiapyrans (11-14) was obtained from examination of the n.m.r. spectra of the dichloro (11) and dithiocyano (13) compounds (Fig. 7 and 8 ) . Conclusions based on a comparison of these 11,

.......

486

L53

=

14, X = COCH,

spectra with that of 1 depended on the assumption t h a t the differences in the electronic distributions in the ground states of the parent system and the substitution products would not radically alter the relative absorptions of the hydrogens common t o both. Thus in the spectrum of 11 (Fig. 7) absorptions with spinspin couplings and relative intensities corresponding to those assigned to the 1-, 3-, and 4-hydrogens of 1 were found within 0.15 p.p.m. of the values for the latter.5 In place of the triplet at 2.58 7 for the 6-hydrogen in 1, however, was a singlet of unit intensity at 2.96 T . The structure of 11 as shown provided the only reasonable interpretation of the spectrum. It was gratifying to note that the quadruplet previously postulated for the 3-hydrogen in 1 but obscured in the spectrum of that compound5 was clearly revealed in the spectrum of (17) T h e corresponding reaction with azulene gave a 97970 conversion of t h e diacetoxymercuri compound t o t h e dithiocyano derivative (18) I t was subsequently discovered t h a t stannic chloride caused t h e polymeriiation of 1. This and t h e fact t h a t t h e presence of an acid and a nucleophile a t t h e same time could brin:. about decomposition (see above) probably account for t h e low yield.

.-

. . .

-

.-..

11. The n.m.r. spectrum of 13 (Fig. 8) was also readily interpreted only in terms of 5,7-substitution. The absorption envelope was similar t o that of 11 but shifted ca. 0.8 p.p.m. t o lower field. This shift probably reflects the large inductive effect of the two thiocyano groups.1g Also the singlet from the 6-hydrogen was superimposed on the downfield portion of the absorption from the 3-hydrogen, and the splitting attributed t o coupIing between the 1- and 3-hydrogens was less than found for 11. These data for 11 and 13 and the analogous mode of formation make structures 12 and 14 quite certain also. The infrared spectrum of 14 showed a carbonyl absorption a t 6.137 M , an even longer wave length than that observed (8.103 F ) for 1,3-diacetylazulene. This indicates relatively more single bond character in the carbonyl groups of 14, and thus that structures such as 15 contribute to the ground state of 14 more than the analogous structures do in the case of 1,3-diacetylazulene. This also suggests that the conjugate acid of 1 would be more stable relative t o 1 than the azulenium ion is relative t o azulene and, therefore, that 1 would be a stronger base than azulene.

15

Br

x

...........

Fig. 8.-S.m.r. spectrum of 5,7-dithiocyanocyclopenta[c]thiapyran in partially deuterated chloroform with tetrainethylsilane a s the internal reference

x = c1

12, X

..

I_._-_-

I

Only a monosubstitution product (16) was obtained in 55% yield from the reaction of 1 and tetranitromethane. The properties of this compound did not provide a definitive distinction between 5- and 7-substitution, and extension of the arguments (above) favoring protonation a t the 7-position to the nitration process provided the only basis of preference. Attempts t o form a dinitro derivative with either tetranitromethane or nitric acid-acetic anhydride as the reagent were unsuccessful. Similarly, treatment of 1 with benzenediazonium fluoroborate gave only a monophenylazo derivative (17) in 787, yield.

9) 9) 5-or

S

/

/ 7-NOZ 16

S /

/

5-or 7-N=NC&&

17

It was of interest t o examine the effects of the different substituent groups on the visible spectrum of 1 and compare these with the data from the corresponding azulene compounds It was found that l 6 and its substitution products displayed the same broad absorption curve with a number of low intensity maxima (19) E G Rordwell a n d P J Bontan J A m Chtvn Soc

7 8 , 8 5 4 (1956)

712

ARTHURG. ANDERSON, JR.,

AND

(poorly resolved in the spectra of 13 and 16) on the long wave length side. The maximum a t the longest wave length was the most distinct and was therefore used for comparisons. As is the case with azulene derivativesz0 the position of this peak varied markedly with the nature of the substituent present (Table I ) . LVhile the direction of the shift caused by a particular group was the same for derivatives of azulene and of 1, the magnitude of the shift in each direction was uniformly less for the derivatives of 1. The specific effects on the visible spectrum of different types of groups in the 1- and 3-positions of azulene have been successfully interpreted2’ with the aid of the values calculated by Pariser for reversal of the direction of the dipole and the consequent changes in the ring atom electron densities accompanying excitation. The results in Table I, TABLE I LONGESTWAVE LENGTH MAXIMA FOR 7- AND CYCLOPENTA [ c ] THIAPYRANS

5 - , and or 7 substituent

Amax, mpa

Ahrnax,

mwh

5,7-SUBSTITUTED

Corresp. azulene --compoundbmBX,A h m a x , Av.,,,, cm.-ib miib cm.-tb

56 5 .. . .. 616 62 -1492 58 -1.544 606 41 -1197 45 -1241 527 -38 1276 -60 1990 NO* 543 -22 717 -48 1556 n-Hexane solvent. This absorption was in most cases not resolved in polar solvents such as methanol and ether. h (or Y ) of parent system - X (or V ) of substituted compound.

Sone CI, c1 Br, Br COCH3,COCHa

and also the hypsochromic shift in the ultraviolet and visible spectra with increasing solvent polarity observed earlier,j are indicative of similar characteristics for 1 : a dipole in the ground state oriented approximately along the long axis of the molecule (18), and a marked decrease, or even a reversal of the direction, of the dipole in the excited state (19) with a consequent decrease in the electron densities a t the 5and ?-positions. The fact that the spectral shifts are less with derivatives of 1 than with the corresponding azulenes suggests that in the former there may be a smaller net change in electron distribution involved.

@ ) . L /

18

fyJ /

s /

WILLIAMF. HARRISOK

Vol. 86

reaction with 1,3-dibrornoazulene was examined. The formation of the nitrobromoazulene product was followed spectrophotometrically at 500 mp and the optical density v s . time measured in runs made a t 50° in aqueous dioxane with varying concentrations of nitrite ion (Table 11). The total concentration of ions in all TABLE I1 RATED A T A

FOR

1,3-DIBROMOAZULESEPLUS S I L V E R AND SITRITE IONS

Runa

1.4g-Ib X 102

[sol-]“ X 102

--

0 min.

-

DIWd 125 min. 232 min. 1085 min.

1

1.05 1.05 0.33 0.85 1.23 1.96 1.05 6.95 .33 .90 1 15 1.83 18 75 .33 .65 0.88 1.63 1 05 a All runs at 50” in 63% dioxane-water. Concentrations of silver and nitrite ions expressed in moles/liter. Excess nitrite ions from added sodium nitrite. Optical density a t 500 mp. 2 3

runs was high compared t o the difference in concentration between runs and it was felt that changes in ionic strength would not be critical. The slight deceleration of the rate found with increased nitrite ion concentrations indicated tha.t this ion was not involved in the rate-determining step.2 3 Kinetic studies on the reactions of silver salts with alkyl halides24point to an appreciable degree of carbonium ion character in the transition state. Therefore if a relatively stable positive ion could be formed, the rate-determining step might involve the abstraction of halide ion by silver ion alone. The product would be formed in a subsequent, more rapid combination of the positive ion with the nucleophile. In that the loss of a halide from a 1-haloazulene would form a positive species (20) which might be stabilized through the delocalization of the positive charge, such a mechanism is proposed for the reaction of 1-haloazulenes with appropriate metal salts. The necessity of sufficient stabilization of the positive charge in the transition state serves to

20

i

19

Reaction of 1,3-Dibrornoazulene and 5,7-Dichlorocyclopenta [clthiapyran with Silver Nitrite.-In the course of concurrent studies on certain of the corresponding 1- and 1,3-substituted azulenes i t was discovered that 1,3-dibromoazulene reacted with silver nitrite in ethanol t o give a 637, yield of l-nitro-3bromoazulene, but no dinitroazulene was formed. 2 2 This result was noteworthy in several respects. It was a nucleophilic substitution a t a ring position known t o be the most reactive to electrophilic species, and only one of the brornines was replaced with excess silver nitrite present. Further, the reaction provided a possible route to an unsymmetrically substituted nitrohalo derivative of 1. To gain some indication of the nature of the reaction process involved, the kinetic dependence on the concentration of nitrite ion of the ( 2 0 ) A. G Anderson, J r , R . G Anderson, and T . S. Fujita, J . O v g . C h e m . , 27, 4.535 ( 1 9 6 2 ) ; A G Anderion, J r , and R : G. Anderson, ibid , 27, 3578 ( 1 9 6 2 ) , and references therein ( 2 1 ) A. G Anderson. J r . , and B M . Steckler, J . A m . Chem. S O L ,81, 4 9 4 1 (1959) ‘ 2 2 ) Subsequent t o the completion of our experiments, similar results were reported for the reaction of 1,3-dibromoazulenederivatives with cuprous cyanide and silver thiocyanate: K Hafner, H. Patzelt and H. Kaiser, Aim.. 686, 24 (1982).

explain the nonreactivity of the 3-nitro- 1-haloazulene products, and of analogous 3-cyano- and 3-thiocyano- 1halo compounds obtained by Hafner, et aLZ2 This interpretation is also applicable t o the reaction of certain halophenols with silver nitrite to form nitrophenols.25 An “ion-pair” mechanism involving “molecular’’ silver salt24is not excluded, however, and further studies in progress include the investigation of this possibility, and also the existence of unpaired electrons in 20. The analogous reaction of 5,7-dichlorocyclopenta[clthiapyran (11) with silver nitrite was found to occur and gave a nitrochloro derivative (21) in 46% yield. (23) Preliminary experiments showed the reaction t o have a first-order dependence on silver nitrite a n d i t was assumed t h a t there was no specific dependence on “molecular” silver nitrite. D m was not determined in the kinetic runs (Table 11) and therefore no k-values were calculated (24) G. S Hammond, M . F Hawthorne, J. H. Waters, and R . M Grayhill, J . A m . C h e m . S o c . , 82, 704 (1960); X Kornblum, R . A . Smiley, R . K . Blackwood. and D. C IMand, i b i d . , 77, 6269 (1965) (25) M . Ringeissen, C o m p f . r e n d . , 198, 2180 (1934)!

Feb. 20, 1964

CYCLOPENTA [CITHIAPYRAN AND 2-PHENYL-2-PYRINDINE

A mechanism involving an intermediate ion (e.g., 22) is also attractive for this case. What was probably the same compound (the ultraviolet and visible spectra, the melting points, and mixture melting point were essentially the same but the infrared spectra, though very similar, were not completely superimposable) was obtained from the reaction of the mononitro derivative (16) with N-chlorosuccinimide. Whether the product was the 5-chloro-7-nitr0, or the 5-nitro-7-chloro compound was not determined. As mentioned earlier, the evidence available slightly favored the 7-position for mononitration and this would lead to the 5-chloro-7nitro structure.

1 2

L.

21

22

The obtaining of this unsymmetrically substituted derivative afforded an opportunity for a rough quantitative test of the additivity of the effects of two unlike groups on the long wave length absorption maximum in the visible region.26 The spectral shift for a chloro group in the 5- or 7-position was assumed t o be onehalf of the amount found for the 5,7-dichloro compound, The calculated shift for 21 on this basis was $4 mp and this was very close to the observed value of $ 3 w.

Experimental2' &Butyl Alcohol-d.-Deuterium oxide (40 9 . ) was equilibrated with dry t-butyl alcohol (74 9 . ) a s described by Albin,28and 40 g. of product was recovered. T h e amount of exchange of the &H hydrogen for deuterium was 95.4 i 1% as determined by the ratio of the areas of t h e peaks for 0 - H absorption in the n.m.r. spectra of t h e deuterated and undeuterated material. 1,2,3,5,7-Pentadeuterioazulene.-Potassium metal (1.5 g., 38.4 mg.-atoms) was dissolved in 20 ml. of absolute t-butyl alcohol (70% deuterated) and 4.5 ml. (ca. 8.6 mmoles of potassium tbutoxide) of the solution thus formed was added to 87 mg. (0.68 mmole) of azulene in a thick walled glass tube. T h e sealed tube and contents were heated a t 120-130" for 18 hr. At the end of this time a n appreciable quantity of decomposition products had separated. T h e tube was cooled, opened, and 99.5yG deuterium oxide was added t o precipitate the azulene. T h e whole mixture was extracted with several portions of ether and the combined ether extracts were then washed with water and dried over sodium sulfate. After filtration, the solvent was removed under reduced pressure and a solution of the residue in n-pentane was chromatographed on Florisil with the same solvent a s eluent. T h e percentage of deuterium exchange a t t h e various positions of the azulene (28 m g . ) obtained from the blue eluate fraction was determined from the ratio of the respective peak areas in the ( 2 6 ) As cyclopenta~clthiapyranis a nonalternant molecule which is iso=-electronic with azulene, and as its spectral properties i n the ultraviolet and visible regions closely parallel those of azulene, it might be expected that the remarkable additivity of the spectral shifts of individual groups found in poly-, and especially in 1,3-substituted azulenes (Plattner's Rule), would hold for the analogous 5,7-substituted heterocyclic analog. For references pertaining to this characteristic of azulenes see PI. A. Plattner. A. Fdrst, and K . Jirasek, H e h Chim Acta. 3 0 , 1320 (1947) A. J . HaagenSmit, Forlschr C h e m . Ovg. A'olurstoje, 6, 40 (1948), ref. 20. For a discussion of the interpretation of this phenomenon in terms of the nonsymmetrical spacing of the energy levels as indicated by simple MO throry see E. Heilbronner, "Non-Benzenoid Aromatic Compounds," ed. by 11. Ginsburg, lnterscience Publishers, l n c . , New York, N Y , 1969, pp 231-24.5, and references therein (27) Melting points are corrected and boiling points are uncorrected unless otherwise indicated. Infrared spectra were recorded with a PerkinElmer Model 21 spectrophotometer with a sodium chloride or calcium fluoride prism. Ultraviolet and visible spectra were taken on a Cary Model 1 I S or 14 spectrophotometer. Suclear magnetic resonance spectra were recorded by Mr. B . J . Nist on a 60-Mc. Varian Associates spectrophotometer, Model V-K3507. Mass spectra were taken by Mr B. J . Nist on a Type 21-103 Consolidated Engineering Corp mass spectrometer. Microanalyses were performed by Drs. G . Weiler and F.B. Strauss, Oxford, England, Dr. Aldred Bernhardt, Max Planck Institute, Miilheim, Germany, or by B . J . Nist, L. H o , A. Kuo. and C . H. Ludwig. (28) J. R. Albin, P h . D . Thesis, University of Washington. 1958.

713

n.m.r. spectra of the product and normal azulene recorded a t the same concentration. T h e values found were I- and 3- ( e a . 68%), 2- j53%),. 5- and 7- (64%,)., and 4-, 6-, and 8- (ea. O%).'O 1,3-Dideuteriocyclopenta[c] thiapyran.-Two milliliters of a solution prepared by the reaction of potassium (0.3 g., 7.38 mg.atoms) in 10 ml. of 95YGdeuterated t-butyl alcohol was added t o 68 mg. (0.508 mmole) of cyclopenta[c]thiapyranin a thick walled glass tube. The sealed tube and contents were kept at room temperature for 3 hr. The tube was opened and the organic material nhich precipitated upon the addition of 99.5% deuterium oxide was collected by filtration and washed with deuterium oxide. A solution of the wet product in 5 ml. of ether was dried over sodium sulfate. Removal of the solvent under reduced pressure left 30 mg. (44%) of product, m.p. 91-93°.5 .4nalysis of the n . m . r . spectrum of the product and comparison of the peak areas with the corresponding ones in the spectrum of undeuterated material taken at the same concentration indicated t h a t greater than 90% exchange had occurred a t the 1- and 3-positions (see the Discussion section). All maxima in the ultraviolet and visible absorption spectra were within 3 mp of those recorded for the undeuterated compound. T h e infrared spectra of the two species, however, were noticeably different. Treatment of cyclopenta[c]thiapyran with potassium t-butoxide in t-butyl alcohol at higher temperatures brought about decolorization of the solution and complete loss of the thiapyran compound. 5,7-Dichlorocyclopenta[c] thiapyran ( 11 ).-To a stirred solution of cyclopenta[c]thiapyran (23 mg., 0.172 mmole) in 5 ml. of benzene was added dropwise over a period of 45 min. a solution of N-chlorosuccinimide (45.8 mg., 0.345 mmole) in 8 ml. of benzene, and stirring was continued for 14 hr. a t room temperature. T h e red solution was then decanted onto a short column of alumina. Removal of the solvent under reduced pressure from the red eluate fraction which separated left a solid which was recrystallized twice from aqueous methanol and afforded 23 mg. (667,) of product 1 1 as red needles, m.p. 86-8i". A n-hexane solution showed Xmax in mp (log e ) a t 234 (4.18), 252 (4.15), 260 (4.15), 281 (4.43), 291 (4.56), 329 (3.57), 333 (3.57), 339 (3.56), 356 (3.41), and 497 (3.19) with shoulders a t 586 (2.67), 601 (2.46), and 617 (2.19). The infrared spectrum was recorded. Anal. Calcd. for C ~ H I S C I QC: , 47.31; H, 1.99; Cl, 34.92. Found: C, 47.04; H , 1.96; C1,34.87. 5,7-Dibromocyclopenta [c]thiapyran ( 12).-To a stirred solution of cyclopenta[c]thiapyran (11.3 mg., 0.0843 mmole) in 5 ml. of methylene chloride was added dropwise over a period of 1 h r . a solution of N-bromosuccinimide (29.7 mg., 0.167 mmole) in 9 ml. of methylene chloride. Stirring was continued a t room temperature for an additional 1.5 hr. during which time some green material precipitated. T h e red solution was treated as described above for the analogous dichloro compound and recrystallization of the solid thus obtained twice from aqueous methanol gave 14.3 mg. ( 5 3 7 ~ of ) product 12 as red needles, m . p . 84-86' dec. A n-hexane solution showed A,, in mp (log e ) a t 232 (4.15), 253 (4.04), 261 (4.07), 283 (4.41), 293 (4.53), 330 (3.48), 339 (3.47), 357 (3.40), and 498 (3.20) with shoulders a t 582 ( 2 . 6 7 ) , 597 (2.48), and 608 (2.19). The infrared spectrum was recorded. Anal. Calcd. for CsH4SBrl: C , 32.90; H , 1.38. Found: C , 32.95; H, 1.62. 5,7-Dithiocyanocyclopenta[c] thiapyran (13).-A 2%, solution of bromine in carbon tetrachloride was added slowly to a cold (O"), stirred suspension of lead thiocyanate (0.113 g., 0.35 mmole) in 2 ml. of carbon tetrachloride until the color of bromine just persisted. Sufficient lead thiocyanate to remove the excess bromine was then added. T h e liquid decantate from the above mixture was added over a period of 5 min. t o an ice-cold, stirred solution of cyclopenta[c]thiapyran (20.3 mg., 0.151 mmole). Stirring was continued for 2 hr. a t 0" and then 30 min. a t room temperature during which time some orange material precipitated. T h e supernatant orange solution was passed through a short column of neutral alumina and the orange fraction which was eluted with methylene chloride was collected. Removal of the solvent under reduced pressure and recrystallization of the residual solid twice from aqueous methanol yielded 18 mg. (48%) of product 13 as orange needles, m.p. 152-184". An ether solution showed, , ,X in mp (log e ) a t 230 (4.25), 244 (4.20), 288 (4.401, 320 (3.64), and 453 (3.49). With n-hexane a s the solvent and the use of 10-cm. cells, the peak in the visible region was shifted to 462 mp b u t no other maxima were resolved. The infrared spectrum (potassium bromide pellet) showed a sharp peak a t 4.65 p which was attributed t o the thiocyano groups. Anal. Calcd. for CLaH4S3N2:C, 48.36; H , 1.62. Found: C , 48.11; H, 1.28. 1,3-Dithiocyanoazulene .-A solution of thiocyanogen in carbon tetrachloride, prepared as described above except t h a t the initial amount of lead thiocyanate was 115.4 mg., was added to a cold (0') stirred suspension of 71 mg. (0.11 mmole) of 1,3-diacetoxy-

714

ARTHURG . ANDERSON,JR.,

niercuriazuleneZ9 in 2 ml. of carbon tetrachloride and the mixture was stirred for 2 hr. a t 0" and then 2 hr. a t room temperature. Treatment of the solution a s described above for the isolation of 5,7-dithiocyanocyclopenta[c]thiapyran afforded 25.9 mg. ( 9 7 q ) of rose-red needles, m.p. 141-144', from the methylene chloride eluate which were identical with an authentic ~ a m p l e . 3 ~ 5,7-Diacetylcyclopenta [c]thiapyran ( 14).-To a solution of cyclopenta[c]thiapyran (28.6 mg., 0.214 mmole) in 7 ml. of acetic anhydride was added stannic chloride (0.36 ml., 3.1 tnmoles) and the mixture was allowed t o stand a t room temperature for 1 hr. The green solution which resulted was poured into a mixture of 77 ml. of 2 'V sodium hydroxide and 50 g . of ice and the whole was shaken for 5 min. and then extracted with successive portions of ether until the extracts were colorless. The solvent was removed from the combined, dried (sodium sulfate) extracts under reduced pressure and the residual red oil was fractionally distilled onto the cold finger of a small sublimation apparatus a t r a . 90" and 1 m m . The solid fraction was twice recrystallized from ether-hexane and gave 4 mg. (9VC)of product 14 as orange needles, m.p. 183184.5'. .I n-hexane solution showed A,, in inp (log e ) a t 240 (4.25), 284 (4.67), 294 (4.43), 308 (4.38), 358 (3.46), 366 (3.53), and 470 (3.64), with shoulders a t 508 (3.391, 517 (3.29), and 527 (3.09). The infrared spectrum (chloroform, calcium fluoride prism) showed strong absorption which was attributed t o the cdrbonyl groups a t 6.136 p. The carbonyl absorption of 1,3diacetylazulene under the same conditions was observed a t 6.103 p . Anal. Calcd. for CL2Hj002S:C, 66.03; H , 4.62. Found: C , 66.14; H , 4.54. 5- (or 7-) Nitrocyclopenta[c]thiapyran (16).-To a stirred solution of cyclopenta[c]thiapyran (16.3 mg., 0.12 mmole) in 0 . 5 ml. of pyridine and 10 ml. of ethanol was added 0.6 ml. of a 0.25 Ai solution of tetranitromethane in ethanol. T h e solution darkened and black material precipitated. After standing 10 min. a t room temperature the mixture was concentrated almost t o dryness under reduced pressure. The residue was taken up in benzene and the orange solution which resulted was passed through a short Florisil column. Elution with methylene chloride, removal of t h e solvent from the orange fraction under reduced pressure, and recrystallization of the residual solid twice from aqueous methanol gave 12 mg. (55 , ) of product 16 as orange needles, m . p . 147.5-150'. An ether olution showed A,,, in mp (log e ) a t 223 (4.13), 249 (4.19), 259 (4.17), 321 (4.23), 350 (3.81), 3.64 (3.81,), 390 ( 3 . 5 3 ) , and 478 (3.72). With nhexane as the solvent in 10-cm. cells shoulders were observed a t 503, 5 2 2 , 532, and 543 mp. Anal. Calcd. for C8HjNS02: C , 53.03; H , 2.79. Found: C , 53.38; H, 2.88. 5 - (or 7-) Phenylazocyclopenta[c]thiapyran ( 17).-To phenyldiazonium fluoroborate (29 mg., 0.115 mmole) was added a cold ( 0 ' ) solution of cyclopenta[c]thiapyran (20 mg., 0.149 mmole) and sodium acetate (60 mg., 0.75 mmole) in 5 ml. of methanol. The resulting violet solution was stirred for 20 min. a t 0 " . The solvent was removed under reduced pressure and the residue was triturated with a few milliliters of methylene chloride. T h e colored solution thus obtained was chromatographed on a short column of neutral alumina with methylene chloride a s the eluent. Removal of the solvent from the violet fraction and recrystallization of the residual solid twice from hexane gave 28 mg. (78'7,) of product 17 as violet prisms, m.p. 94-96', A n-hexane solution showed A,, in mp (log e ) a t 234 (4.16), 290 (4.19), 297 (4.19), 396 (4.21), and 525 (3.83). ;1nai. Calcd. for CI4HIOSN~:C , 70.56; H , 4.23; N , 11.76. Found: C , 69.73; H , 4 . 4 9 ; N , 11.70. (29) D. J Gale, P h . D . Thesis, University of Washington, 1957 ( 3 0 ) A . G . Anderson, Jr , and R. N McDonald. .I. A m C k e m S O L ,81, 366 9 (1959)

AND

WILLIAMF. HARRISON

Vol. 86

1-Nitro-3-bromoazulene .-A mixture of 1,3-dibromoa~ulene3~ (30.7 mg., 0.107 mmo1e)and silver nitrite(30.7 mg., 0.199 inmole) suspended in 8 ml. of ethanol and 3 ml. of water was heated under reflux for 18 hr. and then evaporated to dryness under reduced pressure. A solution of the residue in a little benzene was chromatographed on a column of neutral alumina. Elution with benzene removed unchanged 1,3-dibromoazulene (6.8 mg.); removal of the solvent from the colored fraction eluted with methylene chloride and recrystallization of the residual solid from methylene chloride-n-hexane gave 13.5 mg. (63"; ) of 1nitro-3-bromoazulene, m.p. 184.5187' (lit.31 183-184") which was identical (ultraviolet, visible, and infrared spectra) with an authentic sample. Kinetic Dependence on Nitrite Ion of Reaction of 1,3-Dibromoazulene with Silver Nitrite .--Sufficient 1,3-dibronioazulene was dissolved in 51) ml. of 50c;7, aqueous dioxane to give a spectroscopically determined ( A 630 mp, e 400) concentration of 1.25 X 10-3 M. 'Three solutions were prepared as follows. Each contained 30 Ing. of silver nitrite in 10 nil. of 75':; dioxane-water arid thus LL d v e r ion concentration of 2.1 X 1 0 - 2 .21. Solution A had io other ingredients arid thus was 2.1 X 10-2 .tii i i nitrite ion. Solution B contained 100 mg. of sodium nitrite in addition and was 13.9 X 1 0 P :If in nitrite ion. Solution C contained 300 rng. of sodium nitrite in addition and was 37.9 X 10-2 Ji in nitrite ion. Kinetic runs were made by combining 5 . 0 1711. of the dibroinoazulene solution with 5.0-ml. portions of solutions A , or B, or C. The temperatures of the solutions prior t o mixing and of the reaction mixture were kept a t 50' in a constant temperature bath. Portions of t h e separate runs were withdrawn a t the same time intervals and cooled quickly in ice water. The optical densities a t 500 nip (Table 11) were determined on a Cary spectrophotometer.27 D values were not measured and therefore no k-values were calculated. 5- (7-) Chloro-7- (5-) nitrocyclopenta[c]thiapyran (21).-A mixture of 5,7-dichlorocyclopenta[c]thiapyran(30 mg., 0.149 mrnole) and silver nitrite (130 mg., 0.845 mmole) suspended in 8 ml. of ethanol and 3 ml. of water was heated under reflux for 15 hr. and the solvent was then removed under reduced pressure from the orange solution which resulted. The residue was treated with a few milliliters of methylene chloride and the solution was cliromatographed on a column of neutral alumina with the same solvent as the eluent. Removal of the solvent from. the major orange fraction and recrystallization of the residual solid twice from aqueous methanol gave 14.5 mg. (46'3) of product 21 a s orange-red needies, m.p. 191.5-193". An ether solution showed A, in mp (log e ) a t 244 (4.07), 256 (4.11), 264 (4.12), 332 (4.%3),360 (3.70), 375 (3.72), 400 ( 3 . 5 7 ) , and 495 (3.62). With n-hexane a s the solvent, additional maxima were observed a t 545, 556, and 568 mp. T h e infrared spectrum was recorded. Anal. Calcd. for C8H4C102SS: C , 48.36; H, 1.62. Found: C, 48.11; H, 1.28. Reaction of 7- ( 5 - ? ) Nitrocyclopenta [c] thiapyran with N-Chlorosuccinimide.-To a stirred solution of 7-( 5-?)nitrocyclopenta[clthiapyran (6.5 mg., 0.037 mmole) in 5 ml. of benzene was added slowly a sulution of S-chlorosuccinimide ( 6 mg., 0.045 rnmole) in 5 ml. of benzene. The mixture was stirred a t room temperature for 5 hr. and then chromatographed on a ihort column of neutral alumina. Elution with methylene cliloride gave an orange fraction which, after removal of the solvent under reduced pressure, yielded 6.5 mg. of orange needles, m.p. 165190'. Recrystallization five times from aqueous methanol gave 1.5 mg. (17%) of product, m.p. 189-195", which was indicated t o be the same a s t h a t ( 2 1 ) from the reaction of 5,7-dichloro penta[c]thiapyran with silver nitrite (above) by a mixture m point (189-193") and the qualitative identity of the ultraviolet and visible absorption spectra of the two compounds. The infrared spectra, however, though very similar, were not completely superimposable. (31) A. G Anderson, Jr , J . A. Nelson, and J J T a z u m a , i b i d . , 76, 4980 (1953)